[SOLVED] Burned EPS 8 pin questions

ShowbizAtol933

Honorable
Nov 2, 2015
28
0
10,540
Hello all, after doing some overclocking on my 3960x at 4.8ghz everything seemed stable and working. My motheBroad is an Asus P9X79 pro. Running 1.45v on a Corsair h110i gtx. Suddenly a couple days later my computer would hard shut off under load. So I thought the psu was dying and bought another one. My old one was a Thermaltake 750w smart and my new one is the grey Corsair CX750m. Upon taking off my psu cable from the cpu connector I noticed that my 8 pin was severely burnt. All 4 top power pins were charred. The computer still shut down under stock settings after I turned it back down. Was it my power supply that would cause this or would it be something intenal with the motherboard. I don’t want to plug in my new psu just to have it melt onto the motherboard again and ruin it. As of taking my old psu off the computer would still turn on and idle just fine so I know the cpu isn’t dead.

View: https://imgur.com/nmgjezd
 
Solution
If it was just 1 pin, I'd say a connection issue, too much voltage creating too much heat on a loose pin. But with all 4 12v fried like that, loose connections become serious improbabilities and sub-par connection on the pins becomes more likely. Each of those pins is rated at about 5Amps at 12v DC, so that's 240w on the EPS alone, plus whatever the mains supplies which is at least another 150w. Even with OC you aren't going to overdraw both combined wattage. However, if the wiring isn't large enough, resistance goes up, and so does the amperage to compensate. Added to a cheap crimp at the pin connection, each of those pins was probably seeing right at 5Amps or better at the crimp, and superheating the socket.

What caused the failure...
That is almost certainly an issue with your old power supply, and you should not use that system without a new MOBO and PSU.
What voltage were you pushing on that OC?

Regardless, its amazing your system didnt burn to the ground given that level of damage.
 

ShowbizAtol933

Honorable
Nov 2, 2015
28
0
10,540
That is almost certainly an issue with your old power supply, and you should not use that system without a new MOBO and PSU.
What voltage were you pushing on that OC?

Regardless, its amazing your system didnt burn to the ground given that level of damage.

I was pushing about 1.45v idle and LCC “high” and under HW monitor would state 1.47v under prime testing. Temps on the cpu never got above 80c on the hottest core with a 23c ambient. I also had a fan zipped tied pointed at the VRMs.
 

Karadjgne

Titan
Ambassador
If it was just 1 pin, I'd say a connection issue, too much voltage creating too much heat on a loose pin. But with all 4 12v fried like that, loose connections become serious improbabilities and sub-par connection on the pins becomes more likely. Each of those pins is rated at about 5Amps at 12v DC, so that's 240w on the EPS alone, plus whatever the mains supplies which is at least another 150w. Even with OC you aren't going to overdraw both combined wattage. However, if the wiring isn't large enough, resistance goes up, and so does the amperage to compensate. Added to a cheap crimp at the pin connection, each of those pins was probably seeing right at 5Amps or better at the crimp, and superheating the socket.

What caused the failure would be the increased load under OC. The underlying reason for the failure being a junky psu wire.
 
Solution

ShowbizAtol933

Honorable
Nov 2, 2015
28
0
10,540
If it was just 1 pin, I'd say a connection issue, too much voltage creating too much heat on a loose pin. But with all 4 12v fried like that, loose connections become serious improbabilities and sub-par connection on the pins becomes more likely. Each of those pins is rated at about 5Amps at 12v DC, so that's 240w on the EPS alone, plus whatever the mains supplies which is at least another 150w. Even with OC you aren't going to overdraw both combined wattage. However, if the wiring isn't large enough, resistance goes up, and so does the amperage to compensate. Added to a cheap crimp at the pin connection, each of those pins was probably seeing right at 5Amps or better at the crimp, and superheating the socket.

What caused the failure would be the increased load under OC. The underlying reason for the failure being a junky psu wire.

Damn, I thought Thermaltake psus were decent. So if I had a higher quality psu this could have been potentially prevented?
 

Karadjgne

Titan
Ambassador
Yep, potentially is a good word. TT has a lot of decent stuff, it's got its nose into just about everything thats an accessory for a pc. Fans are decent, AIO's are decent, even their watercooling stuff is decent. But I can't really remark on anything excellent. On the other hand, their psus often leave plenty to be desired in the way of quality. Decent stock replacements, but not even close to Delta or Seasonic or Flextronics built stuff.

From facts observed and some known, you got my best, educated guess but without proof of cause, it's still just an educated guess that happens to fit.

Things like this happening are just another reason why anyone in the Heralds, Moderators and more involved posters will be more than happy to tell you why a Quality psu is the single most important component of the pc. It literally has to power everything else with clean, healthy, correct power and skimping on that can lead to results you got.
 

Karadjgne

Titan
Ambassador
View: https://youtu.be/f6snWfd1v7M


Low end psus, including brand names well recognised, but makes 2 good points. First being that what's claimed on the sticker isn't necessarily true (higher quality psus get reviewed in depth by professionals who put the psu under the microscope so to speak) and second, the complete lack of protections offered by lower quality psus stand a very good chance of passing that damage onto anything downstream, most normally the motherboard and/or gpu but can easily include storage or the cpu or ram or any combination.